Mixed alkyl/aryl phosphonates identify metabolic serine hydrolases as antimalarial targets

Abstract

Malaria, caused by Plasmodium falciparum, remains a significant health burden. One major barrier for developing antimalarial drugs is the ability of the parasite to rapidly generate resistance. We previously demonstrated that salinipostin A (SalA), a natural product, potently kills parasites by inhibiting multiple lipid metabolizing serine hydrolases, a mechanism that results in a low propensity for resistance. Given the difficulty of employing natural products as therapeutic agents, we synthesized a small library of lipidic mixed alkyl/aryl phosphonates as bioisosteres of SalA. Two constitutional isomers exhibited divergent antiparasitic potencies that enabled the identification of therapeutically relevant targets. The active compound kills parasites through a mechanism that is distinct from both SalA and the pan-lipase inhibitor orlistat and shows synergistic killing with orlistat. Our compound induces only weak resistance, attributable to mutations in a single protein involved in multidrug resistance. These data suggest that mixed alkyl/aryl phosphonates are promising, synthetically tractable antimalarials.

Keywords: Plasmodium falciparum; activity-based probes; alky/aryl phosphonates; covalent probes; drug resistance; lipid metabolism; serine hydrolases.

Figure 1:. Structure of Salinipostin A and Synthesized Phosphonates

(A) Structure of Salinipostin A and General Mixed Alkyl/Aryl Phosphonate. Salinipostin A (left) is drawn with electrophilic enolphosphate colored red and the mixed alkyl/aryl phosphonate (right) with electrophilic phosphonate group in red. (B) Synthesis of mixed alkyl/aryl phosphonates and structures of molecules. PPh₃, triphenyl phosphine; I₂, iodine; DCM, dichloromethane; P(OEt)₃, triethyl phosphite; Tf₂O, triflate anhydride; pyr., pyridine; ROH, phenol. (C) Structures of Short Chain Hexyne-based Phosphonates (D)Structures of Pentadecane-based Phosphonates

Figure 2:. 3,5 difluorophenol Substituted Long Chain Phosphonate is a potent inhibitor of parasite growth.

(A) Strucutre of Salinipostin A and constitutional isomers of long chain mixed alkyl/aryl phosphonates. (B) IC₅₀ curves of short chain mixed alkyl/aryl phosphonates for inhibition of 4-methylumbelliferyl caprylate processing by recombinant PfMAGL. (mean ± SD, N,n = 2,3). (C) EC₅₀ values of Compound 21 and 22 in a 72 h treatment of synchronized ring stage W2 parasites with error bars, SEM. (mean ± SD, n = 4 from 3 independent experiments). (D) Rate of kill of Compound 22 measured against compared against known fast (chloroquine), medium (pyrimethamine), and slow (atovaquone) acting inhibitors of parasite growth in which compounds were dosed at their EC₅₀ and parasite viability was monitored over time and ± SEM. N,n = 6,2. (E) HFF Toxicity for alkyl cyclic peptide inhibitor and linear counterparts. Points are plotted as mean ± SEM values of each inhibitor for HFF growth. Data were generated with 72 h assays (N, n = 2,3).

Figure 3:. Chemoproteomics identifies distinct serine hydrolases between active and inactive difluorophenol isomers.

(A) Structures of Activity-Based Probes used for chemoproteomics. (B) Workflow for Competitive ABPP experiment. Infected red bloods are either pretreated with DMSO or an inhibitor (pink) and followed by treatment with an activity-based probe (yellow), before being submitted through typical proteomics workflows. (C) Volcano plots of Activity-based probe 21-alk targets in P. falciparum. The x-axis shows the logarithm values of between 21-alk-treated and DMSO vehicle-treated parasites. The statistical significance was determined without correcting for multiple comparisons in triplicates and the y-value in the volcano plot is the negative logarithm of the p value. Significantly enriched serine hydrolases are denoted in red for parasite enzymes and blue for human enzymes. (D) Volcano plots of Activity-based probe 22-alk targets in P. falciparum. The x-axis shows the logarithm values of between 22-alk-treated and DMSO vehicle-treated parasites. (E) Venn Diagram showing overlap of significantly enriched serine hydrolases between the inactive probe 21-alk red, the active probe 22-alk, and the natural product Salinipostin A.

Figure 4:. Conditional knockdown of abH112 does not lead to a growth defect.

(A) Partial Dendrogram of P. falciparum, Toxoplasma gondii, and human serine hydrolases that share a protein family domain generated using the Clustal Omega algorithm with abH112 and hAB17A colored in red. (B) Overlay of predicted AlphaFold structure of abH112 (orange) and hAB17A (blue) with active site catalytic triad in cyan. (C) Sequence alignment of abH112 and hABHD17A. Lipid binding motif is colored magenta, GXSXG serine hydrolase motif is shown in gold, and catalytic triad is shown in cyan. (D) Western blotting of abH112 cKD lines grown in +/− aTC. Data were measured by ChemiDoc^™ MP Imaging System (Biorad), with bip used as a loading control. (E) Plot of abH112 expression levels from (D) normalized to Bip. (F) Plot of growth levels of parent WT cells (NF54) and abH112 cKD parasites in the presence or absence of aTC. Means are plotted ± SEM (n = 3). Tests for significance used a two-way-ANOVA. G) Plot of EC₅₀ values of parasites for compound 22 in the WT parent cells (NF54) and in the cKD lines in the presence and absence of aTc. Parasites were maintained in 30 nM aTc for 72 hours before treatment with compound 22. Knockdowns were then induced at time of treatment with compound 22 by varying the concentration of aTc. EC₅₀s are plotted as means ± SEMs (N, n = 4–5, 2) and statistical significance determined using Mann-Whitney U tests comparing the parental data vs the abH112 cKD line where *p < 0.05.

Figure 5:. Compound 22 has a mechanism of action different from the pan lipase inhibitor Orlistat.

(A) Giemsa-stained parasites. Synchronized ring-stage parasites were treated at 0–6 h post-invasion with DMSO (vehicle control), 10 µM compound 21 or 22, and imaged at 24 h, 36 h, and 48 h. (B) Lipidomic analysis of compound 22 treated parasites. Synchronized ring stage parasites were treated with either 900 nM compound 22 or DMSO for 24 h before lysis and collection of lipids. Significantly changed lipid classes are denoted on volcano plot. N= 3–4 (C) Isobolograms depicting P. falciparum in vitro culture-based fractional EC₅₀ (FIC50) values in the presence of compound 22 and Orlistat, each point represents a different biological replicate done in technical duplicate.

Figure 6:. Cross resistance and in vitro selection of parasites.

(A) Plot of EC₅₀ values for two SalA resistant lines (harboring mutations in PRELI domain-containing protein at positions P102L and T145I) for compound 22. Values are plotted as means ± SEMs and statistical significance determined using Mann-Whitney U tests where *p < 0.05. (B) EC₅₀ shifts of mutant and parental lines against 22. EC₅₀s for each line are presented as means ± SEMs (N, n = 4, 2) (C) Summary table of EC₅₀s for each line are presented as means ± SEMs (N, n = 4, 2).